Selection of small SNP panels to predict dairy breed proportions of African crossbred cattle

Crosses between exotic dairy cattle and indigenous populations have long been used in African countries to improve productivity. High-grade dairy crosses, however, are superior only if kept under optimal environments. The suboptimal environments of the majority of smallholder farms are better suited for crossbreds with lower exotic breed proportions, with the ideal composition depending on the environment. But most farmers do not know the breed composition of their crossbreds to make appropriate mating decisions. High density genetic markers provide accurate estimates of breed composition. However, for routine application genotyping costs need to be low, which restricts the number of markers that can be genotyped. We determined the accuracy of small SNP assays with 100 to 1,500 markers for estimating total dairy breed proportion of 3,139 crossbred cows from East and West Africa. Markers were selected from around 40k SNPs based on the largest absolute allele frequency differences between pure ancestral breeds. In general, panels with larger proportions of markers differentiating African Bos taurus breeds from exotic taurine dairy breeds performed better compared to panels where most markers were selected to distinguish Bos indicus from exotic dairy breed proportions. Overall, markers selected for largest allele frequency difference between West African indigenous breeds and exotic dairy breeds performed best or near best across all crossbred populations irrespective of their origin with the range of accuracy of prediction ( r2 ) of breed composition being 0.895 to 0.966 (200 SNPs), 0.917 to 0.974 (400 SNPs), and 0.962 to 0.990 (1500 SNPs), respectively.